src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwCSTab_GeomP1P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "vec.h"
  47 #include "nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            CubicSplineTable
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwCSTab_GeomP1P1_VF_c
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     int              i_shift_offset,i_coord_offset,j_coord_offset;
  67     int              j_index_start,j_index_end;
  68     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  69     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  70     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  71     real             *shiftvec,*fshift,*x,*f;
  72     int              vdwioffset0;
  73     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  74     int              vdwjidx0;
  75     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  76     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  77     real             velec,felec,velecsum,facel,crf,krf,krf2;
  78     real             *charge;
  79     int              nvdwtype;
  80     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  81     int              *vdwtype;
  82     real             *vdwparam;
  83     int              vfitab;
  84     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  85     real             *vftab;
  86     int              ewitab;
  87     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  88     real             *ewtab;
  89
  90     x                = xx[0];
  91     f                = ff[0];
  92
  93     nri              = nlist->nri;
  94     iinr             = nlist->iinr;
  95     jindex           = nlist->jindex;
  96     jjnr             = nlist->jjnr;
  97     shiftidx         = nlist->shift;
  98     gid              = nlist->gid;
  99     shiftvec         = fr->shift_vec[0];
 100     fshift           = fr->fshift[0];
 101     facel            = fr->epsfac;
 102     charge           = mdatoms->chargeA;
 103     nvdwtype         = fr->ntype;
 104     vdwparam         = fr->nbfp;
 105     vdwtype          = mdatoms->typeA;
 106
 107     vftab            = kernel_data->table_vdw->data;
 108     vftabscale       = kernel_data->table_vdw->scale;
 109
 110     sh_ewald         = fr->ic->sh_ewald;
 111     ewtab            = fr->ic->tabq_coul_FDV0;
 112     ewtabscale       = fr->ic->tabq_scale;
 113     ewtabhalfspace   = 0.5/ewtabscale;
 114
 115     outeriter        = 0;
 116     inneriter        = 0;
 117
 118     /* Start outer loop over neighborlists */
 119     for(iidx=0; iidx<nri; iidx++)
 120     {
 121         /* Load shift vector for this list */
 122         i_shift_offset   = DIM*shiftidx[iidx];
 123         shX              = shiftvec[i_shift_offset+XX];
 124         shY              = shiftvec[i_shift_offset+YY];
 125         shZ              = shiftvec[i_shift_offset+ZZ];
 126
 127         /* Load limits for loop over neighbors */
 128         j_index_start    = jindex[iidx];
 129         j_index_end      = jindex[iidx+1];
 130
 131         /* Get outer coordinate index */
 132         inr              = iinr[iidx];
 133         i_coord_offset   = DIM*inr;
 134
 135         /* Load i particle coords and add shift vector */
 136         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 137         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 138         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 139
 140         fix0             = 0.0;
 141         fiy0             = 0.0;
 142         fiz0             = 0.0;
 143
 144         /* Load parameters for i particles */
 145         iq0              = facel*charge[inr+0];
 146         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 147
 148         /* Reset potential sums */
 149         velecsum         = 0.0;
 150         vvdwsum          = 0.0;
 151
 152         /* Start inner kernel loop */
 153         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 154         {
 155             /* Get j neighbor index, and coordinate index */
 156             jnr              = jjnr[jidx];
 157             j_coord_offset   = DIM*jnr;
 158
 159             /* load j atom coordinates */
 160             jx0              = x[j_coord_offset+DIM*0+XX];
 161             jy0              = x[j_coord_offset+DIM*0+YY];
 162             jz0              = x[j_coord_offset+DIM*0+ZZ];
 163
 164             /* Calculate displacement vector */
 165             dx00             = ix0 - jx0;
 166             dy00             = iy0 - jy0;
 167             dz00             = iz0 - jz0;
 168
 169             /* Calculate squared distance and things based on it */
 170             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 171
 172             rinv00           = gmx_invsqrt(rsq00);
 173
 174             rinvsq00         = rinv00*rinv00;
 175
 176             /* Load parameters for j particles */
 177             jq0              = charge[jnr+0];
 178             vdwjidx0         = 2*vdwtype[jnr+0];
 179
 180             /**************************
 181              * CALCULATE INTERACTIONS *
 182              **************************/
 183
 184             r00              = rsq00*rinv00;
 185
 186             qq00             = iq0*jq0;
 187             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 188             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 189
 190             /* Calculate table index by multiplying r with table scale and truncate to integer */
 191             rt               = r00*vftabscale;
 192             vfitab           = rt;
 193             vfeps            = rt-vfitab;
 194             vfitab           = 2*4*vfitab;
 195
 196             /* EWALD ELECTROSTATICS */
 197
 198             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 199             ewrt             = r00*ewtabscale;
 200             ewitab           = ewrt;
 201             eweps            = ewrt-ewitab;
 202             ewitab           = 4*ewitab;
 203             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 204             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 205             felec            = qq00*rinv00*(rinvsq00-felec);
 206
 207             /* CUBIC SPLINE TABLE DISPERSION */
 208             vfitab          += 0;
 209             Y                = vftab[vfitab];
 210             F                = vftab[vfitab+1];
 211             Geps             = vfeps*vftab[vfitab+2];
 212             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 213             Fp               = F+Geps+Heps2;
 214             VV               = Y+vfeps*Fp;
 215             vvdw6            = c6_00*VV;
 216             FF               = Fp+Geps+2.0*Heps2;
 217             fvdw6            = c6_00*FF;
 218
 219             /* CUBIC SPLINE TABLE REPULSION */
 220             Y                = vftab[vfitab+4];
 221             F                = vftab[vfitab+5];
 222             Geps             = vfeps*vftab[vfitab+6];
 223             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 224             Fp               = F+Geps+Heps2;
 225             VV               = Y+vfeps*Fp;
 226             vvdw12           = c12_00*VV;
 227             FF               = Fp+Geps+2.0*Heps2;
 228             fvdw12           = c12_00*FF;
 229             vvdw             = vvdw12+vvdw6;
 230             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 231
 232             /* Update potential sums from outer loop */
 233             velecsum        += velec;
 234             vvdwsum         += vvdw;
 235
 236             fscal            = felec+fvdw;
 237
 238             /* Calculate temporary vectorial force */
 239             tx               = fscal*dx00;
 240             ty               = fscal*dy00;
 241             tz               = fscal*dz00;
 242
 243             /* Update vectorial force */
 244             fix0            += tx;
 245             fiy0            += ty;
 246             fiz0            += tz;
 247             f[j_coord_offset+DIM*0+XX] -= tx;
 248             f[j_coord_offset+DIM*0+YY] -= ty;
 249             f[j_coord_offset+DIM*0+ZZ] -= tz;
 250
 251             /* Inner loop uses 74 flops */
 252         }
 253         /* End of innermost loop */
 254
 255         tx = ty = tz = 0;
 256         f[i_coord_offset+DIM*0+XX] += fix0;
 257         f[i_coord_offset+DIM*0+YY] += fiy0;
 258         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 259         tx                         += fix0;
 260         ty                         += fiy0;
 261         tz                         += fiz0;
 262         fshift[i_shift_offset+XX]  += tx;
 263         fshift[i_shift_offset+YY]  += ty;
 264         fshift[i_shift_offset+ZZ]  += tz;
 265
 266         ggid                        = gid[iidx];
 267         /* Update potential energies */
 268         kernel_data->energygrp_elec[ggid] += velecsum;
 269         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 270
 271         /* Increment number of inner iterations */
 272         inneriter                  += j_index_end - j_index_start;
 273
 274         /* Outer loop uses 15 flops */
 275     }
 276
 277     /* Increment number of outer iterations */
 278     outeriter        += nri;
 279
 280     /* Update outer/inner flops */
 281
 282     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*74);
 283 }
 284 /*
 285  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomP1P1_F_c
 286  * Electrostatics interaction: Ewald
 287  * VdW interaction:            CubicSplineTable
 288  * Geometry:                   Particle-Particle
 289  * Calculate force/pot:        Force
 290  */
 291 void
 292 nb_kernel_ElecEw_VdwCSTab_GeomP1P1_F_c
 293                     (t_nblist                    * gmx_restrict       nlist,
 294                      rvec                        * gmx_restrict          xx,
 295                      rvec                        * gmx_restrict          ff,
 296                      t_forcerec                  * gmx_restrict          fr,
 297                      t_mdatoms                   * gmx_restrict     mdatoms,
 298                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 299                      t_nrnb                      * gmx_restrict        nrnb)
 300 {
 301     int              i_shift_offset,i_coord_offset,j_coord_offset;
 302     int              j_index_start,j_index_end;
 303     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 304     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 305     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 306     real             *shiftvec,*fshift,*x,*f;
 307     int              vdwioffset0;
 308     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 309     int              vdwjidx0;
 310     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 311     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 312     real             velec,felec,velecsum,facel,crf,krf,krf2;
 313     real             *charge;
 314     int              nvdwtype;
 315     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 316     int              *vdwtype;
 317     real             *vdwparam;
 318     int              vfitab;
 319     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 320     real             *vftab;
 321     int              ewitab;
 322     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 323     real             *ewtab;
 324
 325     x                = xx[0];
 326     f                = ff[0];
 327
 328     nri              = nlist->nri;
 329     iinr             = nlist->iinr;
 330     jindex           = nlist->jindex;
 331     jjnr             = nlist->jjnr;
 332     shiftidx         = nlist->shift;
 333     gid              = nlist->gid;
 334     shiftvec         = fr->shift_vec[0];
 335     fshift           = fr->fshift[0];
 336     facel            = fr->epsfac;
 337     charge           = mdatoms->chargeA;
 338     nvdwtype         = fr->ntype;
 339     vdwparam         = fr->nbfp;
 340     vdwtype          = mdatoms->typeA;
 341
 342     vftab            = kernel_data->table_vdw->data;
 343     vftabscale       = kernel_data->table_vdw->scale;
 344
 345     sh_ewald         = fr->ic->sh_ewald;
 346     ewtab            = fr->ic->tabq_coul_F;
 347     ewtabscale       = fr->ic->tabq_scale;
 348     ewtabhalfspace   = 0.5/ewtabscale;
 349
 350     outeriter        = 0;
 351     inneriter        = 0;
 352
 353     /* Start outer loop over neighborlists */
 354     for(iidx=0; iidx<nri; iidx++)
 355     {
 356         /* Load shift vector for this list */
 357         i_shift_offset   = DIM*shiftidx[iidx];
 358         shX              = shiftvec[i_shift_offset+XX];
 359         shY              = shiftvec[i_shift_offset+YY];
 360         shZ              = shiftvec[i_shift_offset+ZZ];
 361
 362         /* Load limits for loop over neighbors */
 363         j_index_start    = jindex[iidx];
 364         j_index_end      = jindex[iidx+1];
 365
 366         /* Get outer coordinate index */
 367         inr              = iinr[iidx];
 368         i_coord_offset   = DIM*inr;
 369
 370         /* Load i particle coords and add shift vector */
 371         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 372         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 373         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 374
 375         fix0             = 0.0;
 376         fiy0             = 0.0;
 377         fiz0             = 0.0;
 378
 379         /* Load parameters for i particles */
 380         iq0              = facel*charge[inr+0];
 381         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 382
 383         /* Start inner kernel loop */
 384         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 385         {
 386             /* Get j neighbor index, and coordinate index */
 387             jnr              = jjnr[jidx];
 388             j_coord_offset   = DIM*jnr;
 389
 390             /* load j atom coordinates */
 391             jx0              = x[j_coord_offset+DIM*0+XX];
 392             jy0              = x[j_coord_offset+DIM*0+YY];
 393             jz0              = x[j_coord_offset+DIM*0+ZZ];
 394
 395             /* Calculate displacement vector */
 396             dx00             = ix0 - jx0;
 397             dy00             = iy0 - jy0;
 398             dz00             = iz0 - jz0;
 399
 400             /* Calculate squared distance and things based on it */
 401             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 402
 403             rinv00           = gmx_invsqrt(rsq00);
 404
 405             rinvsq00         = rinv00*rinv00;
 406
 407             /* Load parameters for j particles */
 408             jq0              = charge[jnr+0];
 409             vdwjidx0         = 2*vdwtype[jnr+0];
 410
 411             /**************************
 412              * CALCULATE INTERACTIONS *
 413              **************************/
 414
 415             r00              = rsq00*rinv00;
 416
 417             qq00             = iq0*jq0;
 418             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 419             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 420
 421             /* Calculate table index by multiplying r with table scale and truncate to integer */
 422             rt               = r00*vftabscale;
 423             vfitab           = rt;
 424             vfeps            = rt-vfitab;
 425             vfitab           = 2*4*vfitab;
 426
 427             /* EWALD ELECTROSTATICS */
 428
 429             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 430             ewrt             = r00*ewtabscale;
 431             ewitab           = ewrt;
 432             eweps            = ewrt-ewitab;
 433             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 434             felec            = qq00*rinv00*(rinvsq00-felec);
 435
 436             /* CUBIC SPLINE TABLE DISPERSION */
 437             vfitab          += 0;
 438             F                = vftab[vfitab+1];
 439             Geps             = vfeps*vftab[vfitab+2];
 440             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 441             Fp               = F+Geps+Heps2;
 442             FF               = Fp+Geps+2.0*Heps2;
 443             fvdw6            = c6_00*FF;
 444
 445             /* CUBIC SPLINE TABLE REPULSION */
 446             F                = vftab[vfitab+5];
 447             Geps             = vfeps*vftab[vfitab+6];
 448             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 449             Fp               = F+Geps+Heps2;
 450             FF               = Fp+Geps+2.0*Heps2;
 451             fvdw12           = c12_00*FF;
 452             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 453
 454             fscal            = felec+fvdw;
 455
 456             /* Calculate temporary vectorial force */
 457             tx               = fscal*dx00;
 458             ty               = fscal*dy00;
 459             tz               = fscal*dz00;
 460
 461             /* Update vectorial force */
 462             fix0            += tx;
 463             fiy0            += ty;
 464             fiz0            += tz;
 465             f[j_coord_offset+DIM*0+XX] -= tx;
 466             f[j_coord_offset+DIM*0+YY] -= ty;
 467             f[j_coord_offset+DIM*0+ZZ] -= tz;
 468
 469             /* Inner loop uses 59 flops */
 470         }
 471         /* End of innermost loop */
 472
 473         tx = ty = tz = 0;
 474         f[i_coord_offset+DIM*0+XX] += fix0;
 475         f[i_coord_offset+DIM*0+YY] += fiy0;
 476         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 477         tx                         += fix0;
 478         ty                         += fiy0;
 479         tz                         += fiz0;
 480         fshift[i_shift_offset+XX]  += tx;
 481         fshift[i_shift_offset+YY]  += ty;
 482         fshift[i_shift_offset+ZZ]  += tz;
 483
 484         /* Increment number of inner iterations */
 485         inneriter                  += j_index_end - j_index_start;
 486
 487         /* Outer loop uses 13 flops */
 488     }
 489
 490     /* Increment number of outer iterations */
 491     outeriter        += nri;
 492
 493     /* Update outer/inner flops */
 494
 495     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*59);
 496 }